Discovery of dual Axl/VEGF-R2 inhibitors as potential anti-angiogenic and anti-metastatic drugs for cancer chemotherapy

Article abstract of DOI:10.1016/j.bmcl.2017.06.071

Axl tyrosine kinase has been shown to be involved in multiple pathways contributing to tumor development, angiogenesis, and metastasis. High Axl expression has been observed in many human tumors where it appears to confer aggressive tumor behavior. Here we present several series of dual Axl-VEGF-R2 kinase inhibitors based on extensive optimization of an acyl diaminotriazole. It was hypothesized that dual inhibition of these two receptor tyrosine kinases may have a synergistic affect in inhibiting tumor angiogenesis and metastasis. One of these molecules, R916562 showed comparable activity to Sunitinib in two mouse tumor xenograft models and a mouse corneal micropocket model.

Full text of DOI:10.1016/j.bmcl.2017.06.071

Accepted Manuscript
Discovery of Dual Axl/VEGF-R2 Inhibitors as Potential Anti-Angiogenic and
Anti-Metastatic Drugs for Cancer Chemotherapy
Dane Goff, Jing Zhang, Thilo Heckrodt, Jiaxin Yu, Pingyu Ding, Raj Singh,
Sacha Holland, Weiqun Li, Mark Irving
PII:
S0960-894X(17)30685-6
DOI:
http://dx.doi.org/10.1016/j.bmcl.2017.06.071
BMCL 25105
Reference:
Bioorganic & Medicinal Chemistry Letters
To appear in:
Received Date:
Revised Date:
Accepted Date:
18 April 2017
24 June 2017
26 June 2017
Please cite this article as: Goff, D., Zhang, J., Heckrodt, T., Yu, J., Ding, P., Singh, R., Holland, S., Li, W., Irving,
M., Discovery of Dual Axl/VEGF-R2 Inhibitors as Potential Anti-Angiogenic and Anti-Metastatic Drugs for Cancer
Chemotherapy, Bioorganic & Medicinal Chemistry Letters (2017), doi: http://dx.doi.org/10.1016/j.bmcl.
2017.06.071
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Discovery of Dual Axl/VEGF-R2 Inhibitors as Potential Anti-
Angiogenic and Anti-Metastatic Drugs for Cancer Chemotherapy
Dane Goffa*, Jing Zhanga, Thilo Heckrodta, Jiaxin Yua, Pingyu Dinga, Raj Singha, Sacha
Hollanda, Weiqun Lia, Mark Irvinga
aRigel Pharmaceuticals, 1180 Veterans Blvd., South San Francisco, CA., U.S.A., 94080
*Corresponding author. Current address: 77 Markham Ave., Redwood City, CA, U.S.A.,
94063. E-mail address: dagoff@att.net
Keywords: Axl tyrosine kinase inhibitor, VEGF-R2, diaminotriazole
Abstract
Axl tyrosine kinase has been shown to be involved in multiple pathways contributing to
tumor development, angiogenesis, and metastasis. High Axl expression has been
observed in many human tumors where it appears to confer aggressive tumor behaviour.
Here we present several series of dual Axl-VEGF-R2 kinase inhibitors based on
extensive optimization of an acyl diaminotriazole. It was hypothesized that dual
inhibition of these two receptor tyrosine kinases may have a synergistic affect in
inhibiting tumor angiogenesis and metastasis. One of these molecules, R916562 showed
comparable activity to Sunitinib in two mouse tumor xenograft models and a mouse
corneal micropocket model.
Axl is a receptor tyrosine kinase (RTK) originally identified from patients with
chronic myelogenous leukemia.1 By use of retroviral screening technology, Axl was
shown to regulate endothelial cell migration.2 Axl is also necessary for in vivo
angiogenesis and tumor development in mouse models.2 Axl stimulation triggers several
signaling pathways involved in cellular migration, invasiveness, transformation, and
proliferation as well as angiogenesis and resistance to chemotherapeutic agents.3, 4 High
Axl expression is observed in many human tumors where it appears to confer aggressive
tumor behavior, leading to tumor dissemination and mortality from metastasis.5 We

therefore sought to further validate and characterize Axl as a therapeutic oncology target
by the development of small molecule inhibitors.6
We describe the development of a series of dual Axl-VEGF-R2 kinase inhibitors
culminating in R916562, a diaminotriazole which exhibits antitumor activity comparable
to Sunitinib in a mouse xenograft model. Initial goals were to achieve good Axl potency
as measured by an in-cell western blot (ICWB) assay. Selectivity over the insulin
receptor (INSR) and the closely related RTK Mer were desired.7,8 Non specific anti-
proliferative or cytotoxic activity (PAD assay) was undesirable.9 Compounds were also
screened against vascular endothelial growth factor receptor 2 (VEGF-R2) due to the
known antiangiogenic and antitumor effects of targeting this receptor.10 We hypothesized
that dual Axl/VEGF inhibitors might be very effective oncology therapeutics since
multiple biological pathways involved with angiogenesis and tumorogenesis would be
targeted.
An initial high throughput screen generated an acyl diaminotriazole as a
starting point. Lead optimization led to compound 1. The donor-acceptor-donor triad
provided by the diaminotriazole as well as a basic nitrogen optimally spaced 4-5 atoms
away from the left-hand side aryl ring were essential to the pharmacophore (Fig. 1).
Figure 1 D.A. Goff

Compound 1 showed a 47 nM cellular EC50 against Axl, 2.75 µM against INSR and
1 µM cytotoxicity. Unfortunately 1 had poor aqueous solubility and <1% oral
bioavailability in the SD rat making it unsuitable for testing in a xenograft model.
A logical next step was to replace the acyl carbonyl with a pyridyl nitrogen (Fig1).
The dashed line in Fig, 1 indicates a potential intramolecular hydrogen bond between the
triazole amino group and the pendant carbonyl group of compound 1 or pyridyl nitrogen
of compound 2. Kinase inhibitors designed to contain pseudo six membered rings formed
by intramolecular hydrogen bonding have been reported.11 Indeed, this hypothesis proved
correct as replacing the 4-morpholinophenylcarboxamide of compound 1 with a 2-
pyridyl gave compound 2 which showed an Axl EC50 of 334 nM and a comparable
selectivity profile. Since synthesis of a series of substituted 2-pyridyl analogues did not
result in improved Axl potency, a series of bicyclic analogues was prepared.
Table 1. D.A. Goff
Axl (ICWB) Axl (in-vitro) INSR (ICWB)
Cytotoxicity
Compound Ar
R
EC50(µM)
IC50(µM)
EC50(µM)
EC50(µM)
H
0.004
0.003
0.014
0.040
3
4
H
0.012
0.002
0.582
0.040
H
H
0.022
0.023
0.003
0.014
0.126
1.005
0.040
0.040
5
6

H
H
0.027
0.028
0.007
0.004
0.241
0.300
0.040
0.026
7
8
H
0.052
0.062
0.032
0.016
0.320
4.17
0.040
1.44
9
Cl
10
A variety of quinazoline or fused pyrimidine-substituted diaminotriazoles showed
sub-100 nM inhibition of Axl (Table 1). Diaminotriazoles similarly substituted with
quinolines, isoquinolines and benzothiazoles also showed potent Axl activity (data not
shown), but generally exhibited potent cytotoxicity and INSR activity. An exception was
compound 10, which contained a 6, 7-dimethoxyquinazoline on the right hand side. Data
for some analogues containing the 6,7-dimethoxyquinazoline moiety is shown in Table 2.

Table 2. D.A. Goff
EC50 (ICWB, µM)
Cpd R1
11
R2 Axl
VEGF INSR
H
H
H
0.016
0.039
0.052
0.014
0.003
0.004
1.32
4.58
1.80
12
13
14
15
F
0.055
0.026
0.047
5.03
>10
Cl 0.068
16
H
0.193
0.019
4.76
The desired potency and selectivity for Axl over the INSR and PAD shown by compound
10 was maintained, although Mer selectivity was not achieved. These compounds were
also potent inhibitors of VEGF-R2. Unfortunately, oral rat pharmacokinetics showed high
clearance and volume of distribution for this series.

A breakthrough came when a constrained norbornyl piperazine side chain on the left
hand side was combined with bicyclic heterocycles on the right hand side (Table 3).
Thienopyrimidine 17 showed much-improved clearance (11.5 mL/min/kg), reduced
volume of distribution and a promising half life (4.9 h) in the SD rat. Comparison of
compounds 3 and 21, 4 and 17, and 5 and 18 established that replacement of the left-hand
side 2-pyrrolidinoethoxy group with the norbornylpiperazine significantly reduced
cytotoxicity. Interestingly, compounds 9 and 20 did not show this trend. Most compounds
in the norbornylpiperazine series showed > 50 fold selectivity over INSR but retained
potent VEGFR-2 activity.

Table 3. D.A. Goff
Cpd
15
Ar
Axl (ICWB) Axl (in-vitro) VEGF (ICWB) Cytotoxicity PK
(SD Rat)a
EC50 (µM)
IC50 (µM)
EC50 (µM)
EC50 (µM)
Cl=117, Vss=27
T1/2=2.7h
0.068
0.009
. 0.047
4.18
Cl=11.5, Vss=4.4
T1/2=4.9h
17
18
19
20
21
0.170
0.253
0.350
0.397
0.485
0.009
0.010
0.015
0.032
0.003
0.065
0.146
n.d.
0.56
0.51
2.40
0.04
0.75
Cl=44, Vss=7.9
T1/2=2.3h
Cl=52, Vss=8.8
T1/2=2.3h
0.104
0.196
n.d.b
n.d.
a i.v. clearance in mL/min/kg., Vss=volume of distribution in L/kg
bn.d. means not determined. All compounds are racemic exo isomers
A further series of thienopyrimidine analogues of 17 (Table 4) included many analogues
with improved Axl inhibition. Of these, only thienopyrimidines bearing a 5-methyl
substituent (17 and 26) or a fused aromatic ring (25) had acceptable cytotoxicity. Given

its good pharmacokinetics and acceptable activity and selectivity, compound 17 was
selected for further study.
Table 4 D.A. Goff
EC50 (ICWB, µM)
Cpd Ar
R
Axl
VEGF
Cytotox.
23
24
F
0.020 n.da
0.040
0.051
H
0.021 0.024
25
26
H
F
0.052 0.024
0.062 0.090
1.26
0.500
27
28
Cl
H
0.084 0.080
0.085 0.160
0.049
0.037
The preparation of the two enantiomers of compound 17 is shown in Scheme 1. PCC oxidation
of commercially available (+)-endo-Norborneol followed by reductive amination with
4-nitrophenylpiperazine and catalytic hydrogenation gave (1S, 2S, 4R) 1-bicyclo[2.2.1]
hept-2-yl-4-(4-aminophenyl)piperazine 28. Reaction with diphenyl N-cyanocarbonimidate
in isopropanol gave N-cyano-O-phenyl isourea 29. The preparation of the required
2-chloro-4-hydrazino-7-methylthieno[3, 2-d] pyrimidine 30 started with the conversion
of commercially available methyl 3-amino-4-methylthiophene-2-carboxylate to

2,4-dihydroxy-7-methylthieno[3,2d]pyrimidine 32 12 followed by chlorination and
mono-hydrazinolysis (Scheme 2). Intermediates 29 and 30 were heated in NMP at 135oC
for 15 h to give R916562. On a 40 gram scale this gave 46% yield. The same reaction in
a CEM microwave apparatus for 6 minutes at 180oC gave R916562 in 50% yield on a 1 gm
scale. The preparation of the enantiomer R952417 started with the (1R, 2R, 4S)
N-cyano-O-phenylisourea 31 which was separated by chiral hplc from the racemate.13 By SCF chiral
chromatography R916562 had an e.e. of 95.5% while R952417 had an e.e of 98.5%.
Scheme 1 D.A. Goff

Scheme 2 D.A. Goff
The configuration of the norbornyl ring did not greatly affect the cytotoxicity or pharmacokinetics.
Since R916562 showed better cellular activity than 952417 against Axl (136 nM vs. 216 nM) and
could be made from commercially available material, it was selected for study in mouse xenograft models.
R916562 was tested in subcutaneous xenograft models in nude mice using two tumor
cell lines.14 In the MDA-MB-231 human breast cancer xenograft model, both tumor
growth delay (TGD) and tumor growth inhibition (TGI) were observed (Figure 2). Oral
dosing with R916562 at 125 mg/kg b.i.d. resulted in 79% TGD. One complete remission
and one partial remission were also observed (n=10). R916562 treatment at 85 mg/kg or
125 mg/kg orally b.i.d for 21 days resulted in statistically significant TGIs of 69% or
83% respectively, comparable to 84% TGI observed for the positive control Sunitinib
given once daily at 80 mg/kg.15 R916562 also showed efficacy in the Caki-1 human renal
carcinoma xenograft model. A statistically significant 80% TGI (p<0.05; n=10) was
achieved using 85 mg/kg b.i.d R916562, comparable to 85% TGI with Sunitinib (80
mg/kg, once daily) in this model.

Figure 2
a.
b.
100
900
Vehicle
Vehicle
90
80
70
60
50
40
30
20
10
0
Sutent 80mg/kg QD
R562 85 mg/kg BID
R562 125 mg/kg BID
untreated
800
700
600
500
400
300
200
100
0
R562 125mg/kg BID
Treatment
0
2
4
6
8
10 12 14 16 18 20 22
0
5
10 15 20 25 30 35 40 45 50 55 60
Day of study
Day of study
Moreover, R916562 showed 73% reduction in fibroblast growth factor–induced
neovascularization in a mouse corneal micropocket assay at a dose of 100 mg/kg and
50% reduction at 50 mg/kg. Both results were highly statistically significant (P<0.001,
Figure 3). In this model the anti-angiogenic effect of R916562 was again comparable to
that of the positive control Sunitinib (78% reduction at 80 mg/kg).
Figure 3
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
80
25
50
100
Dose(mg/kg)
Treatment
Stimulus
-
Vehicle Sutent
R562
bFGF-
bFGF+
*** Statistically significant, Kruskall-Wallace Dunn test,
p<0.001. R562 was dosed bid; Sutent was dosed qd

In summary, a series of potent dual inhibitors of Axl kinase and VEGF-R2 have been
developed around a diaminotriazole core. Off-target INSR activity and cytotoxicity
was reduced to acceptable levels, although selectivity against closely-related TAM family
RTK Mer was not achieved. Several members of this series were suitable for oral
delivery and one of these, R916562, showed activity comparable to the clinically
approved tyrosine kinase inhibitor Sunitinib in mouse tumor xenograft models and the
corneal micropocket angiogenesis model. Interestingly, anti-Axl antibodies have recently
shown only moderate efficacy (40% inhibition) in an A549 tumor cell mouse xenograft
model.16 Inhibition of both Axl and VEGF-R2 might therefore be more effective in
preventing angiogenesis and tumor growth than targeting either receptor alone. In fact we
observed (and it has been subsequently reported) that Sunitinib shows potent 9 nM in-
vitro inhibiton of Axl, as well as potent VEGF-R2, Flt-3 and Kit activity.17 The potential
therapeutic benefits of using dual Axl/VEGF-R2 inhibitors, especially in patients who
have relapsed after VEGF-R2 targeted therapy have been highlighted in a recent
Perspective on Axl inhibitors in cancer.18
Acknowledgement. We would like to thank Yasumichi Hitoshi, Alison Pan, Joanne
Litvak, Catherine Sylvain, Betty Chang, Christian Franci, Allan Torneros, Matt Duan,
Polly Pine and Donald G. Payan for their support.
Supplementary Data
Supplementary data describing the synthesis of R916562 can be found on the online
version at xxxx
References and Notes
1. O’Bryan, J.P.; Frye, R.A.; Cogswell, P.C., Mol. Cell Biol. 1991, 11, 5016.
2. Holland, S.J.; Powell, M.J.; Franci, C.; Chan, E.W.; Friera, A.M.; Atchison, R.E.;
McLaughlin, J.; Swift, S.E.; Pali, E.; Yam, G.; Wong, S.; Lasaga, J.; Shen, M.R.; Yu, S.;
Xu, W.; Hitoshi, Y.; Bogenberger, J.; Nor, J.E.; Payan, D.G.; Lorens, J.B., Cancer Res.
2005, 65, 9294.
3. Hafizi, S.; Dahlback, B., Cytokine Growth Factor Rev. 2006, 17, 295.

4. Linger, R.M.; Keating, A.K.; Earp, H.S.; Graham, D.K., Adv. Cancer Res. 2008, 100,
35.
5. Holland, S.J.; Pan, A.; Franci, C.; Hu, Y.; Chang, B.; Li, W.; Duan, M.; Torneros, A.;
Yu, J.; Heckrodt, T.J.; Zhang, J.; Ding, P.; Apatira, A.; Chua, J.; Brandt, R.; Pine, P.;
Goff, D.A.; Singh, R.; Payan, D.G.; Hitoshi, Y., Cancer Res. 2010, 70, 1544.
6. A portion of this work was presented: Holland, S.J.; Hu, Y.; Franci, C.; Duan, M.;
Bagos, A.; Torneros, A.; McLaughlin, J.; Zhang, J.; Yu, J.; Heckrodt, T.H., Litvak, J.;
Stauffer, E.; Clemens, G.; Daniel-Issakani, S., Pine, P.; Goff, D.A.; Singh, R.; Payan,
D.G.; Hitoshi, Y. Poster #598 AACR-NCI-EORT International Conference, San
Francisco, CA, Oct 22-26, 2007. See also Goff, D.; Zhang, J.; Sylvain, C.; Singh. R;
Holland, S.; Yu, J.; Heckrodt, T.H.; Ding, P., WO 2008083356, 2008.
7. Biochemical activity was determined using recombinant Axl protein. The cell based
Axl activity was determined in HeLa cells using the downstream phosphorylation of Akt
by Axl as a readout. For details see ref.5.
8. Axl is a member of the TAM (Tyro-3, Axl, Mer) receptor tyrosine kinase family.
Inhibition of Mer is potentially undesirable since Mer knockout mice have been shown to
suffer from autoimmune deficiencies and blindness. For reviews see a) Linger, R.M.A.;
Keating, A.K.; Earp, H.; Shelton, G.; Douglas, K., Advances in Cancer Research 2008,
100, 35. b) Lemke, G.; Rothlin, C.V., Nature Rev. Immunology 2008, 8, 327.
9. Catalano, S.M.; McLaughlin, J.; Hitoshi, Y.; Payan, D., Assay and Drug
Development Technologies 2009, 7, 180.
10. Ferrar, N.; Gerber, H.P.; LeCouter, L., Nature Medicine 2003, 9, 685.
11. a) Furet, P.; Caravatti, G.; Guagnano, V.; Lang, M.; Meyer, T.; Schoepfer, J., Biorg.
Med. Chem. Lett. 2008, 18, 897. b) It has been proposed that Sunitinib is capable of
forming a pseudocycle via an intramolecular hydrogen bond. See Hughes, T.V.;
Emanuel, S.L., O’Grady, H.R., Connolly, P.J., Rugg, C.; Fuentes-Pesquera, A.R.;
Karnachi, P.; Alexander, R.; Middleton, S.A., Bioorg. Med. Chem. Lett. 2008, 18, 5130.
12. Jourdan, F.; Laduree, D.; Robba, M., J. Heterocyclic Chem. 1994, 31, 305.
13. Intermediates 29 (retention time= 64.1 min) and 31 (retention time= 61.9 min)
were separated on a Chiralcel-05 column (4.6 x 250 mm) eluting with MeOH containing
0.2 volume % TFA at a flow rate of 0.7 mL/min.

14. The tumor xenograft and corneal micropocket studies were performed at Piedmont
Research Center LLC, Morrisville, N.C. TGD is defined as the percent increase in time to
endpoint of treated versus control mice while TGI is the difference in median tumor
volumes between treated and control mice at day 21. In general a TGI of >60% indicates
potential therapeutic value. At 85 mg/kg b.i.d. R916562 was well tolerated, but at 125
mg/kg some toxicity was observed.
15. Sunitinib is a VEGF-R2 inhibitor clinically approved for the treatment of
gastrointestinal and renal cancer. Review: Adams, V.R.; Leggas, M., Clinical
Therapeutics 2007, 29, 1338.
16. Li, Y.; Ye, X.; Tan, C.; Hongo, J.A.; Zha, J.; Kallop, D.; Ludlam, M.J.C.; Pei, L.,
Oncogene 2009, 1.
17. Karaman, M.W.; Herrgard, S.; Treiber, D.K.; Gallant, P.; Atteridge, C.E.; Campbell,
B.T.; Chan, K.W.; Ciceri, P.; Davis, M.L.; Edeen, P.T.; Faraoni, R.; Floyd, M.; Hunt,
J.P.; Lockhart, D.J.; Milanov, Z.V.; Morrison, M.J.; Pallares, G.; Patel, H.K.; Pritchard,
S.; Wodicka, L.M.; Zarrinkar, P.P., Nat. Biotechnol. 2008, 26, 127.
18. Myers, S.H.; Brunton, V.G.; Unciti-Broceta, A. J. Med. Chem. 2016, 59, 3593.
Captions to Figures and Tables
Figure 1
Evolution of the Acyl Diaminotriazole Series to a 2-Pyridyl Diaminotriazole. The
dashed line shows a potential hydrogen bond to form a pseudo six-membered ring.
Figure 2
R916562 blocks growth of MDA-MB-231 breast carcinomas implanted subcutaneously
in nu/nu mice. a) increase in mean tumor volume- 79% tumor growth delay after 21 days
of treatment; b) percent survival: 1 partial remission and 1 complete remission were
observed after treatment with R916562.
Figure 3
R916562 in the Mouse Corneal Micropocket Model
Scheme 1

Preparation of R916562 and its enantiomer R952417. Reagents and conditions:
(a) PCC, dichloromethane, 0oC; (b) Na(OAc)3BH (1.4 eq.), HOAc (2 eq.), 1,2-
dichloroethane ; (c) 10% Pd/C, MeOH; (d) diphenyl N-cyanocarbonimidate, 2-
propanol; (e) 30 (1.05 eq.), NMP,135-140oC, o.n.

Scheme 2
Preparation of intermediate 30. Reagents and conditions: (a) KOCN (1.6 eq.), HOAc,
r.t.; (b) aq. NaOH, reflux, 1H, cool, then 2M HCl: (c) POCl3, reflux: (d) Hydrazine,
ethanol, reflux.
Table 1
Axl, Insulin Receptor and PAD Activity for bicyclic diaminotriazoles
Table 2
Axl, Insulin Receptor and VEGF-R2 activity for dimethoxyquinazoline substituted
diaminotriazoles
Table 3
SAR and pharmacokinetics of diaminotriazoles substituted with a norbornylpiperazine
Table 4
SAR of diaminotriazoles substituted with a racemic norbornylpiperazine and a
thienopyrimidine

Graphical abstract